Megadeals took 87.5% of US venture dollars, and the rest of the market is priced by vintage
THE SO WHAT
When 87.5% of US venture dollars go into $100M+ rounds, capital for sub-megadeal companies stops being about sector and starts being about when you last raised. If you’re not in the AI megadeal club, plan for a vintage discount on your next round and manage burn like external up-rounds are rare, not assumed.
READ THE SOURCE
MORE FROM THE WIRE
Startups & VentureIndex in talks to lead $500m round for DeepMind researcher’s new AI lab
A $500M first check into a new lab from a DeepMind alum says frontier-scale capital is now comfortable backing talent plus thesis, not just incumbents. If you’re building infra, evals, or tooling, assume another well-funded foundation model competitor is entering your pipeline and adjust your partnership and hiring roadmap accordingly.
Startups & VentureHow a $250 million acquisition collapsed into allegations of fraud and forged signatures
A $250M deal unraveling into fraud and forged-signature claims is a governance warning shot—late-stage AI and software M&A is moving faster than many control environments. If you’re buying or selling, tighten diligence on cap tables, signatures, and board approvals this week; don’t assume standard processes kept up with the last two years of growth.
Startups & VentureSilicon Data, which offers real-time compute pricing data to financial institutions and exchanges, raised a $30.5M Series A led by the Valor Atreides AI Fund
A $30.5M Series A for real-time compute pricing is a tell that GPU capacity is being financialized like a commodity. If your business depends on large-scale training or inference, expect more transparent—and more volatile—pricing; start instrumenting your workloads so you can arbitrage clouds, regions, and vendors instead of taking list price.
Startups & VenturePartnering with Preview: Lights, Inference, Action
Sequoia backing Preview is another data point that inference-time orchestration — not just bigger base models — is where venture dollars see defensible value. If you’re building in applied AI, expect investor diligence to center on how you manage context, tools, and latency at inference, not just your fine-tune story.